Author: Foo Yun Chee and Tassilo Hummel
BRUSSELS (Reuters) – Landmark European rules on artificial intelligence will come into effect next month, setting a potential global tone for the technology used in business and daily life, EU countries on Tuesday approved a political deal struck in December. benchmark.
The EU’s AI bill is more comprehensive than the U.S.’s lax voluntary compliance approach, while China’s approach is aimed at maintaining social stability and state control.
Two months ago, EU lawmakers backed artificial intelligence legislation drafted by the European Commission in 2021, with some key changes.
In recent months, as generative AI systems such as Microsoft-backed OpenAI’s ChatGPT and Google’s (Nasdaq: ) chatbot Gemini have grown in popularity, concerns about artificial intelligence have led to misinformation, fake news and copyright infringement. Concerns about protected materials are growing globally.
Mathieu Michel, Belgium’s Minister of Digitalization, said in a statement: “This landmark law, the first of its kind in the world, addresses global technological challenges while also creating opportunities for our society and economy.
He said: “With the Artificial Intelligence Bill, Europe is underlining the importance of trust, transparency and accountability when dealing with new technologies, while ensuring that this rapidly changing technology can thrive and promote European innovation.”
The AI Bill imposes strict transparency obligations on high-risk AI systems, while such requirements will be less stringent for general-purpose AI models.
It limits the government’s use of instant biometric surveillance in public places to handle certain crimes, prevent terrorist attacks and search people suspected of the most serious crimes.
Patrick van Eecke of law firm Cooley said the impact of the new legislation would extend beyond the 27-nation bloc.
“The bill will have global implications. Companies outside the EU that use EU customer data in their AI platforms will need to comply. Other countries and regions may use the bill as a blueprint, as they did with the GDPR. ” He was referring to EU privacy rules.
While the new regulations will be implemented in 2026, once they come into effect, the use of artificial intelligence in social scoring, predictive policing and the untargeted grabbing of facial images from the internet or CCTV footage will be banned for six months .
Obligations for general-purpose AI models will apply after 12 months, while rules for AI systems embedded in regulated products will apply after 36 months.
Depending on the type of violation, non-compliance fines range from €7.5 million ($8.2 million), or 1.5% of turnover, to €35 million, or 7% of global turnover.
(1 USD = 0.9199 Euro)
